4th down: Taking a look at sub-optimal decisions

Last night, some strange 4th down decisions were made during the Texans-Bengals and the Vikings-Packers game; most egregiously by the Vikings. 

All win probabilities will be courtesy of Brian Burke's 4th down calculator.  Win probabilities when going for it, do NOT assume an automatic conversion.

Texans-Bengals

Case 1:

Houston's 2nd drive in the 1st Quarter, they faced 4th & 2 at the Texans' own 30 and decided to kick a field goal with the score tied at 0.

The calculator says that they have a 56% win probability by going for it, a 54% win probability by kicking a field goal and a 50% win probability by punting.

The calculator assumes an 80% probability of making the field goal; however Houston Kicker Shayne Graham is just 69% for the year beyond the 40 yard line - implying that in this specific situation, the odds are less than NFL norms of a successful field goal attempt albeit through a small sample size.  While Cincy allowed lower than average conversion rates on all 4th downs (41%), Houston has a 57% 4th down conversion rate on the year and the break-even point is 44%.  Given the circumstances, the Texans likely would have been better to go for it - the above figures were ALL 4th downs - most of those 4th down attempts would have likely been late in the game when a team HAD to have it - case in point - the league average is 60% on 4th & 2.

Houston likely had better than a 44% chance of converting. 

Case 2:

With about 8:00 left in the 3rd, Houston faced 4th & 1 at midfield ahead 16-10.  The calculator says that had they gone for it, their win probability was 76% with a break-even point of 62%.  By punting, their win probability was 74%. 

Here, the conversion rates from earlier don't apply - that was all 4th downs - the league average for converting a 4th & 1 is 76%.

This decision is more defensible because we can't use stats in a vacuum.  Cincy had yet to show any signs of life offensively; on the flip side, because points are at a premium, a conversion gives you 1st & 10 in Cincy territory with a chance to make it a 2 possession game.  I think this is debatable either way.

Case 3:

With just over 14:00 left to play in regulation, Cincy found themselves down 19-10 facing 4th & 2 at Houston's 29.  The calculator suggests by going for it, Cincy had a 20% win probability and the field goal attempt gave them an 18% win probability with a break even point of just 41%.

In situations where you are down a TD + a FG, taking the points is tempting for sure.  The problem with taking the field goal is that an opponent field goal makes it a 2 possession game again.   Granted, Houston's no offensive juggernaut and Cincy has a stiff defense which you can argue makes the field goal scenario more tempting because you have good reason to put faith in your defense.

On the flip side if you are the Bengals, you have had trouble moving the ball all day and you need a TD at some point.  Why not try when you are inside the opponent's 30? 

 Packers-Vikings

Case 1:

Trailing 3-0 on their 1st possession facing 4th & 1 on their own 38, the Packers punted.  The calculator suggests that going for it gave them a 41% chance of winning while punting gave them a 38% chance of winning.

Your defense just allowed points on the prior possession and you gave up 37 points in your last encounter with this team - not to mention the fact you have Aaron Rodgers.

Case 2:

Trailing 7-3 midway through the 2nd quarter, the Vikings faced 4th & 1 from their own 23.

Conventional wisdom says, 'if you fail, you are basically giving your opponent points'.  However, statistics tell us that the most likely outcome by going for it improves your win probability from 32% by punting to 34% by going for it - the breakeven point is 55% and the league average conversion rate is 76%.

The Vikings and Packers are not two average teams.  The Vikings have Adrian Peterson on their team and in the two prior encounters with the Packers he rushed for over 400 yards!   This is a man who just rushed for the 2nd most yards in a season EVER - why on earth would you punt in any 4th & short situation? Consider the other side of the ball - Green Bay has Aaron Rodgers and averaged nearly 30 points in your two encounters - there is a reasonable chance they would score anyway without the excellent field position.

Case 3:

On the very next drive following Case 2, the Pack faced 4th & goal from the Vikings' 1.  By kicking the field goal, McCarthy gave up 5 points in win probability (78 by going for it vs. 73 with the field goal). Note:  In most cases, our example coaches merely made sub-optimal choices - not awful choices.  5 points in win probability is a lot.

  
Case 4:

This doesn't impact win probabilities but is more of a general principle.

The Vikings faced 4th & 2 from their own 17 down by 3 touchdowns with 10:53 to play.  Did they go for it?  Nope.  Did they punt?  Nope.  They called a Timeout!

For what purposes?  At this point, your win probability was about 1%.  What good is the timeout going to do?

And if you trail by 3 touchdowns, why punt at all?  To hold down the margin of defeat?   And this is a playoff game - there is no tomorrow.  Again - you have Adrian *cking Peterson, who actually had a reasonably productive game.